Disparities in Health Care and Medical Evaluations by Gender: A Review of Evidence and Mechanisms

Gender gaps are observed in many settings, from educational investments to labor market outcomes. Observed disparities by gender extend to health-care settings, where prior work suggests that women may be evaluated differently or receive fewer treatments than men with similar conditions. Prior work has also documented gender gaps in benefit receipt within social insurance programs that rely on medical evaluations, such as Social Security disability insurance and workers’ compensation insurance.

While this prior work documents important correlations between patients’ gender and medical evaluations/treatments, these correlations are only suggestive, as there may be other factors—unobserved to the econometrician—that are correlated with gender and that may account for the observed differences in treatments and evaluations. Even if unobservable factors are constant across male and female patients, the mechanisms that would lead to differences in evaluation and treatment outcomes are unknown.

This article proceeds as follows. Section I reviews the prior evidence on gender gaps in health care and disability benefit receipt. Section II discusses potential mechanisms underlying these gender gaps. Finally, Section III discusses new evidence on the role of providers in contributing to gender gaps.

I. Literature Review on Gender Gaps in Health Care and Disability

Various studies have shown a correlation between patient gender and treatment among patients with similar health conditions. For example, compared to male patients, female patients are less likely to receive aggressive treatment in response to several different heart-related issues (Hernandez et al. 2007, Pelletier et al. 2004) and are also less likely to receive prophylaxis for blood clots when hospitalized (Pietropaoli et al. 2010). Among patients with irritable bowel syndrome, male patients receive more imaging than female patients do, while female patients are more likely to receive tranquilizers and lifestyle advice than male patients ) (Hamberg, Risberg, and Johansson 2004. Among emergency department patients reporting similar levels of abdominal pain, females are less likely than males to receive treatment to alleviate pain (Chen et al. 2008).

These correlations could arise if physicians interpret symptoms differently for male and female patients. Some evidence suggests that providers often view reports of pain differently for male and female patients. For example, compared to males, females reporting pain are more likely to be told that their pain is emotionally driven rather than arising from a physical impairment (Côté and Coutu 2010, Hoffmann and Tarzian 2001).

How physicians assess health conditions and impairments matters for determining the future health care that patients receive. Further, physicians’ assessments of impairment can also influence access to benefits in social insurance programs.

Consistent with gender differences in impairment evaluations, prior work reveals male-female disparities exist in access to disability benefits from social insurance programs. For example, Low and Pistaferri (2019) show that female applicants for Social Security disability insurance are 20 percentage points more likely to be rejected than similar male applicants, and Card and McCall (2009) document that females’ workers’ compensation claims are 19 percent more likely to be denied than males’ workers’ compensation claims.

II. Potential Mechanisms

The reasons for the differential treatments and outcomes by patient gender are not clear. As male and female bodies are biologically different, the differential treatment and outcomes may arise because of true differences in symptoms or appropriate medical treatment for similar diagnoses. Similarly, women may be more likely than men to apply for disability benefits when they do not meet the criteria. It is possible that men and women communicate different information to providers regarding symptoms and preferences over treatment options.

There are reasons to think that health-care providers could play a role in observed disparities in evaluations and treatment. For example, in a survey by the Kaiser Family Foundation and The Undefeated (Kaiser Family Foundation 2020) about experiences with physicians in the past three years, women were more likely than men to report that a health-care provider had assumed something without asking, talked down to them, didn’t believe them, or refused to order a test or treatment they thought that they needed.

If providers play a role in differential assessments for male and female patients, providers’ gender may matter if male and female physicians differ systematically from each other in their assessments of male and female patients. Some evidence points to patient-provider demographic concordance being relevant for the care that patients receive. For example, recent evidence suggests that racial concordance of patients and doctors affects the health care that Black patients receive, with the take-up of preventive care being higher for Black patients matched with Black doctors than for Black patients matched white doctors (Alsan, Garrick, and Graziani 2019). Other work suggests that racial and gender concordance between providers and patients is associated with reduced mortality in hospital settings (Greenwood, Carnahan, and Huang 2018; Hill, Jones, and Woodworth 2018). In contrast though, physicians assessing hypothetical vignettes have been found to make similar treatment decisions for men and women with similar symptoms (Weisse, Sorum, and Dominguez 2003).

The possibility that male and female health-care providers systematically differ in their assessments and treatment of male and female patients is particularly important in light of the gender distribution of physicians. Despite gender parity in the training of new physicians, nearly two-thirds of current physicians are men (Kaiser Family Foundation 2019). In specialties that often assess physical impairments, like orthopedics, the gender imbalance is even larger (Jagsi et al. 2014). The gender imbalances in the composition of providers could play a role in disparities in impairment assessments if patients and physicians having similar characteristics and backgrounds improves physician-patient communication or reduces physician bias against a patient or against the types of impairments that the patient is likely to have. However, studying the role of provider sex in medical evaluations is difficult because both patients and physicians can often influence patient-physician matches.

Some prior work points to differential sorting of patients to providers based on sex. McDevitt and Roberts (2014) document that urology patients disproportionally select urologists of the same gender. This preference for same-gender providers has the potential to lead to health-care access issues for females, since only 6 percent of US urologists are female despite 30 percent of urology patients being female. McDevitt and Roberts further document that markets with no female urologists have higher death rates for female bladder cancers, suggestive of potential health effects associated with the scarcity of female urologists.

Female patients preferring treatment from a female physician suggests that female patients may feel better about the care they receive from female providers. However, the ability of patients to influence the gender of their providers makes interpreting differences between male and female providers difficult, since differential outcomes could arise from factors related to selection.

III. New Evidence on the Role of Providers

While an emerging literature has documented large gender disparities in evaluations, treatments, and associated social insurance benefit receipt, it is unclear what drives these disparities and what interventions may affect them. It is possible that gender disparities in outcomes may be explained by gender differences in factors unobserved to the researcher but observed to providers. Or, it could be that differences in treatments by gender may be justified if the expected benefits of particular treatments vary by patient gender. Alternatively, it could be that provider discretion or patient-provider communication may be an important factor explaining these disparities. Understanding the underlying sources of these gender disparities is essential for interpreting these disparities and understanding which policies may affect them.

In new work, Cabral and Dillender (2021) leverage a setting with random assignment of doctors to patients to provide the first evidence on the role of providers—and in particular, the role of provider gender—in contributing to gender disparities in medical evaluations. The setting analyzed in Cabral and Dillender (2021) is the Texas workers’ compensation system, which allows insurers and injured workers to request independent medical evaluations to settle disputes over a worker’s level of ongoing disability. Conditional on an injured worker’s county, assignment of the worker to a doctor is random among doctors with similar credentials, and workers’ compensation insurers must pay benefits to injured workers based on the doctor’s assessment. In this setting, all patients seen by doctors believe they suffer from an impairment from a work-related injury that prevents them from working. Randomly assigned doctors are tasked with examining these claimants and assessing whether they are truly impaired and their degree of ongoing impairment.

To analyze how the evaluating doctor’s gender factors into claimants’ subsequent benefits, Cabral and Dillender (2021) investigate how the outcomes of these exams differ based on gender of both providers and claimants using data on exams occurring from 2013 to 2017 and on all health care and income benefits that claimants have received through the Texas workers’ compensation system.

As discussed in Cabral and Dillender (2021), there are several strengths of the setting for evaluating the impact of providers on medical evaluations. First, random assignment of providers to patients allows the researchers to overcome the key identification challenge of potential sorting on unobservables and identify causal effects. Second, the setting includes evaluations of both female and male patients, allowing for separate identification of the provider-patient gender match effect from both provider gender and patient gender effects. Third, the administrative data used in Cabral and Dillender (2021) contain rich baseline demographic and medical information, which can be used to verify that provider sex is uncorrelated with baseline patient characteristics and to explore heterogeneity. Fourth, the data contain information on a large number of randomized medical evaluations, with more than 70,000 exams performed by more than 1,000 doctors. Finally, the outcome of these medical exams is important for the patient, as the assigned doctor makes a binding determination regarding the patient’s subsequent disability benefit receipt.

Cabral and Dillender (2021) find that female patients are less likely to be evaluated as having a continuing disability and thus receive less in subsequent disability benefits when evaluated by male doctors instead of female doctors. In contrast, there is little impact of doctor sex on the disability evaluations and subsequent benefits for male patients.

IV. Conclusion

There is growing evidence of gender disparities in medical evaluations, treatments, and outcomes. Still, little is known about the mechanisms that contribute to these disparities and what policies may affect these disparities.

Qualitative studies and survey evidence indicate that women are more likely to have their symptoms dismissed and more often told that their symptoms are due to emotional sources rather than physical sources. These studies point to the potential importance of provider discretion and provider-patient communication in contributing to gender disparities in medical evaluations and subsequent outcomes. Alignment of patient and provider demographic characteristics may impact provider discretion and patient-provider communication.

A new study by Cabral and Dillender (2021) provides the first evidence on the role of doctors and doctors’ gender in gender disparities in medical evaluations and subsequent disability benefit receipt by investigating a setting where doctors are randomly assigned to patients. The researchers find female patients are assessed as being more disabled and thus subsequently receive substantially higher benefits when evaluated by female doctors rather than male doctors. There is no such relationship between disability evaluations for male patients and the sex of the evaluating doctor. Taken together, the findings in Cabral and Dillender (2021) suggest that increasing the share of female doctors in the provider workforce would increase disability benefits for women and shrink gender disparities in benefit receipt. As many health-care treatments depend on how providers diagnose and evaluate patients, the findings in Cabral and Dillender (2021) highlight the potential importance of gender diversity among health-care professionals in contributing to gender disparities in health-care settings more broadly.